diff --git a/src/Storages/FileLog/FileLogSource.cpp b/src/Storages/FileLog/FileLogSource.cpp
index eb8cb2b28f9..0c266178142 100644
--- a/src/Storages/FileLog/FileLogSource.cpp
+++ b/src/Storages/FileLog/FileLogSource.cpp
@@ -42,6 +42,8 @@ FileLogSource::FileLogSource(
     size_t files_per_stream = file_infos.file_names.size() / max_streams_number;
     start = stream_number * files_per_stream;
     end = stream_number == max_streams_number - 1 ? file_infos.file_names.size() : (stream_number + 1) * files_per_stream;
+
+    storage.increaseStreams();
 }
 
 void FileLogSource::onFinish()
diff --git a/src/Storages/FileLog/FileLogSource.h b/src/Storages/FileLog/FileLogSource.h
index 0ff71c1af87..cdf60cd4fc1 100644
--- a/src/Storages/FileLog/FileLogSource.h
+++ b/src/Storages/FileLog/FileLogSource.h
@@ -34,6 +34,7 @@ public:
     {
         if (!finished)
             onFinish();
+        storage.reduceStreams();
     }
 
 protected:
diff --git a/src/Storages/FileLog/StorageFileLog.cpp b/src/Storages/FileLog/StorageFileLog.cpp
index 6489c9270ca..fb13137ce31 100644
--- a/src/Storages/FileLog/StorageFileLog.cpp
+++ b/src/Storages/FileLog/StorageFileLog.cpp
@@ -42,6 +42,7 @@ namespace ErrorCodes
     extern const int CANNOT_READ_ALL_DATA;
     extern const int LOGICAL_ERROR;
     extern const int TABLE_METADATA_ALREADY_EXISTS;
+    extern const int CANNOT_SELECT;
 }
 
 namespace
@@ -326,6 +327,11 @@ Pipe StorageFileLog::read(
             table_id.getTableName());
     }
 
+    if (running_streams.load(std::memory_order_relaxed))
+    {
+        throw Exception("Another select query is running on this table, need to wait it finish.", ErrorCodes::CANNOT_SELECT);
+    }
+
     /// We need this lock, in case read and streamToViews execute at the same time.
     /// In case of MV attached during reading
     std::lock_guard<std::mutex> lock(status_mutex);
@@ -364,6 +370,16 @@ Pipe StorageFileLog::read(
     return Pipe::unitePipes(std::move(pipes));
 }
 
+void StorageFileLog::increaseStreams()
+{
+    running_streams.fetch_add(1, std::memory_order_relaxed);
+}
+
+void StorageFileLog::reduceStreams()
+{
+    running_streams.fetch_sub(1, std::memory_order_relaxed);
+}
+
 void StorageFileLog::drop()
 {
     try
diff --git a/src/Storages/FileLog/StorageFileLog.h b/src/Storages/FileLog/StorageFileLog.h
index 82c7a2dc9f9..bdab85866c0 100644
--- a/src/Storages/FileLog/StorageFileLog.h
+++ b/src/Storages/FileLog/StorageFileLog.h
@@ -122,6 +122,9 @@ public:
             throw Exception(ErrorCodes::LOGICAL_ERROR, "The key {} doesn't exist.", key);
     }
 
+    void increaseStreams();
+    void reduceStreams();
+
 protected:
     StorageFileLog(
         const StorageID & table_id_,
@@ -169,6 +172,12 @@ private:
     };
     std::shared_ptr<TaskContext> task;
 
+    /// In order to avoid data race, using a naive trick to forbid execute two select
+    /// simultaneously, although read is not useful in this engine. Using an atomic
+    /// variable to records current unfinishing streams, then if have unfinishing streams,
+    /// later select should forbid to execute.
+    std::atomic<int> running_streams = 0;
+
     using TaskThread = BackgroundSchedulePool::TaskHolder;
 
     void loadFiles();